correlog {pgirmess} | R Documentation |
Computes Moran's or Geary's coefficients on distance classes from a set of spatial coordinates and corresponding z values
correlog(coords, z, method="Moran", nbclass = NULL,...)
coords |
a two columns array, data.frame or matrix of spatial coordinates. Column 1 = X, Column 2 = Y. |
z |
a vector for the values at each location. Must have the same length as the row number of coords |
method |
the method used. Must be "Moran" (default) or "Geary" |
nbclass |
number of bins. If NULL Sturges method is used to compute an optimal number |
... |
further arguments to pass to e.g. moran.test or geary.test |
Uses the library spdep including moran.test
or geary.test
. These methods assume the data are normally distributed. Distances are euclidian and in the same unit as the spatial coordinates. Moran's Ho: I values larger than 0 due to chance;
Geary's Ho: C values lesser than 1 due to chance. Correlog has print and plot methods; statistically significant values (p<0.05) are plotted in red.
An object of class "correlog", a matrix including:
class |
bin centers |
I |
the coefficient values |
p.value |
probability of Ho |
n |
the number of pairs |
Computing can take a long time for large data sets
Patrick Giraudoux pgiraudo@univ-fcomte.fr and Colin Beale c.beale@macaulay.ac.uk
see library spdep
library(spdep) data(oldcol) attach(COL.OLD) coords<-cbind(X,Y) res<-correlog(coords,CRIME) plot(res) res<-correlog(coords,CRIME,method="Geary") plot(res)